Sleep Hypoxemia as a Predictor of Mortality in Patients with Sleep Apnea

Background

The Sleep Heart Health Study (SHHS) was a prospective cohort study formulated to explore the risk factors for development of cardiovascular disease in OSA, diagnosed via unaccompanied home sleep apnea test. We used these data to compare the association of the apnea-hypopnea index (AHI) and amount of sleep hypoxemia with the risk of all-cause mortality.

Research Question

What is the relationship among hypoxemia, AHI, and mortality in OSA?

Study Design and Methods

We compared the association of (1) the AHI, (2) the percentage of sleep time with oxygen saturation < 85% (PERC85), and (3) the duration of sleep (in minutes) with oxygen saturation < 85% (MIN85) with the risk of all-cause mortality in the SHHS. Multivariable logistic regression analyses were used and adjusted for age, sex, BMI, pack-years of smoking, cardiovascular score at baseline, and treatment status.

Results

PERC85 was associated with an increased risk of death (OR, 1.03; 95% CI, 1.01-1.05; P = .003). Patients with PERC85 of 1% to 5%, 5% to 20%, and > 20% showed progressively higher risks compared with those with PERC85 of < 1% (1%-5%: OR, 1.37 [95% CI, 1.02-1.83]; 5%-20%: OR, 1.76 [95% CI, 1.07-2.86]; > 20%: OR, 2.93 [95% CI, 1.20-6.98]; P < .05 for all). The MIN85 predicted all-cause mortality (OR, 1.01 [95% CI, 1.00-1.01]; P = .009). Participants with 2 to 30 min and > 30 min of PERC85 showed higher likelihoods of death vs those with PERC85 of < 2 min (2-30 mins: OR, 1.29 [95% CI, 1.01-1.63]; > 30 min: OR, 2.15 [95% CI, 1.22-3.76]; P < .05 for all). AHI was not associated with an increased risk of mortality.

Interpretation

Our findings indicate that sleep hypoxemia with MIN85 is a better predictor of mortality in OSA than AHI. Monitoring oxygen saturation levels and duration may be important for risk stratification and assessment of treatment adequacy in OSA, although this may be confounded by hypoxemia not related to OSA.